OCR vs intelligent document processing comes down to one line. OCR turns an image of a document into text, letters and numbers with no meaning attached, while intelligent document processing, IDP, uses AI to understand what the document is, who sent it and what each field means, and then checks that data against the ERP before it is used.
In the office the difference shows up on the first invoice. OCR gives you the text of the PDF. Someone still reads it, finds the purchase order, compares the lines and types the invoice in. IDP does the reading, the finding and the comparing, which is what supplier invoice capture means when it actually saves time.
What OCR does
Turns pixels into characters
A scan or a PDF image becomes searchable text. On a clean document that works well, and it has for decades.
Works with zones or templates
To know that a number is the total, classic OCR needs to be told where the total is. A template per layout, and a broken template when the layout changes. It does not know which supplier sent the invoice or whether the quantity is right.
What intelligent document processing does
Understands the document
It recognises an invoice, a packing list, a customer order or a certificate, in any layout, in more than one language, without a template per sender.
Identifies the sender
The supplier or the customer behind the document is tied to its record in the ERP, with the open orders.
Checks against the ERP record
Each line is compared with the purchase order and the goods receipt. This is three-way matching, done before entry.
Stops exceptions and enters the rest
A price above the PO, a missing receipt, a duplicate. Each waits for a person with a short explanation. What matches is created in the ERP as a native record, with no retyping.
Where OCR is enough, and where it is not
Archiving and search
If the goal is a searchable archive of scanned paper, OCR does the job. A single form that never changes, from a single source, is fine too.
Many senders, many layouts
Hundreds of suppliers, each with its own invoice. Dozens of customers with their own order forms. Templates cannot keep up. This is the daily reality of intelligent document processing for freight forwarders and for any AP team.
The value is in the check
Reading the price is easy. Knowing it is wrong needs the PO. OCR never sees the PO.
How it works with DOConvert
DOConvert is the AI layer for the ERP. It connects directly to Priority through the DOConvert Priority connector, which is bi-directional, and to Oracle Fusion Cloud ERP and Infor M3, no middleware. Here is one supplier invoice.
The invoice arrives by email
A PDF, a scan or a photo. DOConvert identifies the supplier from the sender and the document, with no template, and reads the open purchase orders from the ERP.
The lines are checked
Quantity and price against the PO, received quantity against the goods receipt. Say one line out of 15 is above the PO price.
The exception stops
The line stops with a plain reason for the buyer. That is the human in the loop at the decision point, and only there.
The invoice is entered
Once decided, the invoice is created in the ERP, linked to the PO and the receipt. Which kind of AI should do this is the debate Avi Rafalson, DOConvert’s CEO, takes up in Generalist vs specialist, the real debate in IDP models.
Who it fits
- Controllers and AP managers who tried OCR and still have people checking and typing
- VPs of Operations with orders, delivery notes and import files arriving by email
- Logistics and freight teams with thousands of documents a month from hundreds of senders
- IT and IS managers on Priority ERP, Oracle Fusion Cloud ERP or Infor M3
Questions and answers
Does IDP use OCR inside?
Often, yes, as the first step. The difference is everything that happens after the text exists.
Do I need a template for each supplier with IDP?
No. A first invoice from a new supplier is read and checked like any other.
Is IDP accurate enough to enter invoices without a person?
The matching lines, yes. Lines that do not match the PO stop for a person.
How long does it take to set up?
A typical DOConvert implementation takes about 10 business days, in the cloud or on your own servers.
Want to see it on your own documents? Book a Demo or Get Started